Dear Sir,

We recognised the problem described in the Brief Report “Intrahepatic right-to-left shunting after the Fontan operation” by Giamberti and co-authors¹ with assent, however want to draw your attention to an alternate operative strategy. The authors understandably express their fear of possible hepatic necrosis subsequent to ligation of hepatic veins, because they were unsure about the adequacy of hepatic venous drainage to the inferior caval vein. Although reports on this complication are scarce, it is probably impossible to get certainty on hepatic venous distribution prior to the secondary intervention.

An additional consideration is that separate hepatic venous drainage to the atrium serve as a substitute for fenestration in the setting of total cavopulmonary connection. After all, the hepatic venous blood that would normally enter the pulmonary circulation, constitutes a right-to-left shunt when it enters the atrium and adds to the preload of the systemic circulation. A consideration of a totally different nature is that this construction – first described by Kawashima² – predisposes to pulmonary arteriovenous fistula, because of an unknown hepatic venous substance that prevents formation of these fistula. Nonetheless, from a haemodynamic standpoint, this “Kawashima operation” has the same effect on systemic preload as a fenestration has, thus to construct both is probably superfluous.

When we were prompted in 1995 to limit hepatic venous flow to the atrium, because of a cavo-atrial shunt via the liver (right-to-left), we banded the hepatic vein, instead of ligating it for fear of hepatic necrosis by venous obstruction³. Banding enabled us to calibrate the right-to-left shunt to a systemic oxygen saturation of ca. 90%, thus keeping systemic preload within a target range and at the same time precluding hepatic venous obstruction. Whether this fear of hepatic venous obstruction is justified, is unknown. We only want to draw attention to the possibility of banding the liver vein in stead of ligating it in the setting of intrahepatic right-to-left shunting after a cavopulmonary connection thus keeping the haemodynamic equivalent of a fenestration. Finally, when the banding is done with non-absorbable material, the relative amount of right-to-left shunt will diminish with the growth of the patient.