Hepatocellular carcinoma is the fifth most common cancer worldwide. Percutaneous interventional therapies, such as radiofrequency (thermal) ablation (RFA), have been developed for early hepatocellular carcinoma. RFA competes with other interventional techniques such as percutaneous ethanol injection, surgical resection, and liver transplantation. The potential benefits and harms of RFA compared with placebo, no intervention, chemotherapy, hepatic resection, liver transplantation, or other interventions are unclear. To assess the beneficial and harmful effects of RFA versus placebo, no intervention, or any other therapeutic approach in patients with hepatocellular carcinoma. We searched the Cochrane Hepato-Biliary Group Controlled Trials Register, the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, and ISI Web of Science to September 2012. We handsearched meeting abstracts from ASCO, ESMO, AASLD, EASL, APASL, and references of articles. We also contacted researchers in the field (last search September 2012). We considered for inclusion randomised clinical trials investigating the effects of RFA versus placebo, no intervention, or any other therapeutic approach on hepatocellular carcinoma patients regardless of blinding, language, and publication status. Two review authors independently performed the selection of trials, assessment of risk of bias, and data extraction. We contacted principal investigators for missing information. We analysed hazard ratios (HR) as relevant effect measures for overall survival, two-year survival, event-free survival, and local recurrences with 95% confidence intervals (CI). In addition, we analysed dichotomous survival outcomes using risk ratios (RR). We used trial sequential analysis to control the risk of random errors ('play of chance'). We identified no trials comparing RFA versus placebo, no intervention, or liver transplantation. We identified and included 11 randomised clinical trials with 1819 participants that included four comparisons: RFA versus hepatic resection (three trials, 578 participants); RFA versus percutaneous ethanol injection (six trials, 1088 participants) including one three-armed trial that also investigated RFA versus acetic acid injection; RFA versus microwave ablation (one trial, 72 participants); and RFA versus laser ablation (one trial, 81 participants). Ten of the eleven included trials reported on the primary outcome of this review, overall survival. Rates of major complications or procedure-related deaths were reported in 10 trials. The overall risk of bias was considered low in five trials and high in six trials. For a subgroup analysis, we included only low risk of bias trials. Regarding the comparison RFA versus hepatic resection, there was moderate-quality evidence from two low risk of bias trials that hepatic resection seems more effective than RFA regarding overall survival (HR 0.56; 95% CI 0.40 to 0.78) and two-year survival (HR 0.38; 95% CI 0.17 to 0.84). However, if we included a third trial with high risk of bias, the difference became insignificant (overall survival: HR 0.71; 95% CI 0.44 to 1.15). With regards to the outcomes event-free survival and local progression, hepatic resection also yielded better results than RFA. However, the number of complications was higher in surgically treated participants (odds ratio (OR) 8.24; 95% CI 2.12 to 31.95). RFA seemed superior to percutaneous ethanol or acetic acid injection regarding overall survival (HR 1.64; 95% CI 1.31 to 2.07). The RR for mortality was also in favour of RFA, but did not reach statistical significance (150/490 (30.6%) people in the percutaneous ethanol or acetic acid group versus 119/496 (24.0%) people in the RFA group; RR 1.76; 95% CI 0.97 to 3.22). The proportion of adverse events did not differ significantly between RFA and percutaneous ethanol or acetic acid injection (HR 0.70; 95% CI 0.33 to 1.48). Trial sequential analyses revealed that the number of participants in the included trials was insufficient and that more trials are needed to assess the effects of RFA versus other interventions. The effects of RFA versus no intervention, chemotherapeutic treatment, or liver transplantation are unknown. We found moderate-quality evidence that hepatic resection is superior to RFA regarding survival. However, RFA might be associated with fewer complications and a shorter hospital stay than hepatic resection. We found moderate-quality evidence showing that RFA seems superior to percutaneous ethanol injection regarding survival. There were too sparse data to recommend or refute ablation achieved by techniques other than RFA. More randomised clinical trials with low risk of bias and low risks of random errors assessing the effect of RFA are needed.