Thirteen trials (2,535 patients) were included in the review comprising two randomised controlled trials (RCTs) and 11 non-RCTs. Trial sample size ranged from 68 to 419 patients. The quality of the included trials was variable: two trials scored 4 points (higher quality); one trial scored 3 points; one trial scored 2 points; and nine trials scored 1 point (low quality). The duration of follow-up ranged from 22.7 to 847 [as given in the paper] months.
Survival rates: Compared with surgical resection, radiofrequency ablation was associated with a statistically significant lower overall survival at one year (OR 0.60, 95% CI 0.42 to 0.86; 12 trials; Ι²=15%). This result remained significant at three and five years (full results presented in the paper). There was no evidence of publication bias with this outcome. Subgroup analysis in tumours 3cm or smaller indicated that radiofrequency ablation was associated with a statistically significant lower overall survival at one year (OR 0.34, 95% CI 0.13 to 0.89; five trials; Ι²=0%) compared with surgical resection; these results remained significant at three and five years.
Recurrence rate: Compared with surgical resection, radiofrequency ablation was associated with a statistically significant increased recurrence rate at one year (OR 1.48, 95% CI 1.05 to 2.08; 13 trials; Ι²=63%). This result remained significant at three and five years (full results presented in the paper). There was no significant difference in local recurrence. There was no evidence of publication bias with recurrence.
Complication rate: Compared with surgical resection, radiofrequency ablation was associated with a statistically significant increased complication rate (OR 6.25, 95% CI 3.12 to 12.52; seven trials; Ι²=54%). There was no evidence of publication bias with this outcome.
Sensitivity analysis reportedly indicated that the results were robust.