A physiopathological interpretation is offered of the possibilities given by recent selective or portally regulated reduction of pressure in achieving a balance between reduction of pressure and diversion. Classic demonstrations, in fact, prove effective in the prevention of haemorrhage due to rupture of oesophageal varices. On the other hand, by over-diversion of the portal flow from the liver, they may cause a further deterioration in liver performance and more imposing the greater the flow of portal blood through the liver preoperatively. Procedures aimed at de-tension of the portal system, particularly the gastro-oesophageal segment, and at ensuring an intrahepatic venous flow sufficient to obviate steal syndrome damage, are described. Some of these procedures are still at the experimental stage. The techniques involved are explained, along with the underlying principles achieved by or prospected for each of them.