Band ligation for primary prevention of bleeding from oesophageal varices is better than no treatment, except when the varices are small, say the authors of a Cochrane review.

Their meta-analysis examining the use of ligation for primary prevention of upper gastrointestinal bleeding concluded that further trials would be considered unethical.

The study, which included six randomised controlled trials conducted between 1996 and 2005, found that 71 of 320 people with liver cirrhosis, but no history of upper GI bleeding, who underwent ligation died, compared with 129 of 317 participants who did not undergo the procedure.

This equated to a reduction in mortality risk of 56% that meant only six patients needed to be treated to avoid one death, they said.