Randomised controlled trials for variceal bleeding

The variceal bleeding episode represents several days of high risk of bleeding, thus therapy should be evaluated not only in terms of immediate cessation of bleeding but also in terms of providing a bleed-free interval of a few days. As the risk of continued bleeding or very early rebleeding from varices diminishes rapidly following admission, time is an important confounding variable when comparing therapies within and between trials. Cirrhotics with better liver function are more likely to stop bleeding with simple measures than those with worse liver function. Vasopressin, glypressin, vasopressin combined with nitroglycerin and somatostatin have all been used as splanchnic arteriolar vasoconstrictors thus reducing portal pressure. No trial has demonstrated increased survival with use of these agents. The efficacy of vasopressin is now disputed. Vasopressin combined with nitroglycerin and somatostatin have the fewest side-effects and may be more effective than vasopressin alone. Balloon tamponade arrests bleeding and prevents exsanguination, but should be used solely as a temporizing measure before the use of emergency sclerotherapy or surgery. Sclerotherapy is the only non-operative emergency technique which has been shown not only to stop variceal bleeding, but to reduce the frequency of very early rebleeding. Emergency oesophageal transection is equally if not more effective in arresting bleeding than sclerotherapy and has a lower early rebleeding rate and a similar mortality. Choice of treatment depends on expertise available. Further studies in the management of variceal bleeding should evaluate 3 main areas. Firstly improvement of existing therapies or new therapies. Secondly, investigation of therapies not related to bleeding, eg prophylaxis against infection, improvement in renal support. Lastly evaluation of predictive factors which may a priori determine a high risk of continued bleeding or early rebleeding thus justifying immediate sclerotherapy or surgery in a sub-group of patients.