Both treatment approaches were equally effective but the intraoperative ERCP group had less morbidity, a shorter hospital stay, and reduced costs. The lower morbidity in the intraoperative ERCP group resulted from the lower rate of papillotomy and lower rates of post-ERCP pancreatitis and cholecystitis. Total morbidity was principally related to the type of treatment approach used.

Radioactive microspheres were used to determine the hepatic haemodynamic response to portacaval anastomosis in normal, cirrhotic and chronic prehepatic portally hypertensive rats 20 days after operation, and in normal rats 2 months after operation. After 20 days portacaval anastomosis caused a decrease in liver mass only in normal and cirrhotic animals, whereas hepatic arterial blood flow per unit of mass increased in normal (+488 per cent), cirrhotic (+191 per cent) and prehepatic portally hypertensive rats (+133 per cent). Despite these facts, animals with portacaval anastomosis showed a reduced hepatic total perfusion (arterial plus portal inflow) per unit of mass with respect to controls in normal (-53 per cent) and cirrhotic rats (-68 per cent), but not in those with prehepatic portal hypertension. Comparing studies carried out at 2 months with those performed 20 days after portacaval anastomosis in normal rats, some recovery of liver mass and total liver blood flow was observed. In conclusion, portacaval anastomosis produced a limited increase in hepatic arterial blood flow which was unable to preserve liver mass and its total perfusion in normal and cirrhotic animals. In contrast, portacaval anastomosis did not significantly alter liver mass or its perfusion in animals with chronic prehepatic portal hypertension, as both values were previously diminished in controls. Thus, the risk of liver failure after portacaval anastomosis is higher in normal and cirrhotic rats than in those with chronic prehepatic portal hypertension.

The effect of surgical end-to-side portacaval anastomosis (PCSA) on systemic and splanchnic circulation has been studied in cirrhotic rats with portal hypertension (CCl4-phenobarbital method) and in control animals. Hemodynamics have been measured using the microsphere technique, with a reference sample for the systemic hemodynamic measurements, and intrasplenic injection for portal systemic shunting rate measurements. Compared with controls, sham-operated (SO) cirrhotic rats showed a hyperdynamic circulation with increased cardiac output (CO) and decreased mean arterial pressure and peripheral resistances. PCSA in control rats induced only a small change in systemic hemodynamics, with parallel decreases in arterial pressure and peripheral resistances, and a small, nonsignificant increase in CO. In cirrhotic rats, PCSA induced a decrease of CO to values similar to those of control rats, with an increase in total peripheral resistances. PCSA induced an increase in hepatic arterial blood flow in control and in cirrhotic rats, portal pressure becoming in this latter group not different from that of control rats. Blood flow to splanchnic organs was higher in SO cirrhotic than in SO control animals. Thus portal venous inflow was also increased in SO cirrhotic rats. PCSA induced an increase in portal venous inflow in control rats, which was only significant in cirrhotic rats when expressed as a percentage of CO. In SO control animals, a significant correlation was observed between total peripheral resistances and splanchnic arteriolar resistances and between CO and splanchnic blood flow. These correlations were not observed in cirrhotic rats. These results do not support the hypothesis that hyperdynamic circulation shown by cirrhotic rats is based on increases in splanchnic blood flow and (or) massive portal systemic shunting.

Case reportA 35-year-old woman presented with upper gastrointestinal haemorrhage in September 1989. Four years earlier, chronic active hepatitis (HbAg+) had been diagnosed by percutaneous liver biopsy, the procedure being remembered as especially painful. She had no previous history of gastrointestinal bleeding, blood transfusion, alcoholism or drug addiction. Physical examination revealed moderate ascites and splenomegaly, although the blood biochemical determinations, coagulation studies and leucocyte and platelet counts were normal. Endoscopy showed large bleeding oesophageal varices which were successfully sclerosed. Ultrasonography demonstrated a central anechoic lesion in the right lobe of the liver which was interpreted as a cyst. The patient rebled from oesophageal varices 1 month later and endoscopic variceal sclerotherapy was again successful. Portohepatic echo-Doppler examination revealed hepatofugal blood flow. Splanchnic and hepatic angiography (October 1989) demonstrated a hypertrophied and tortuous hepatic artery feeding a very active intrahepatic pseudoaneurysmal arterioportal fistula located centrally in the right lobe, with immediate retrograde filling of the portal vein ( Figure I ) and reflux to the left gastric vein, with extensive collateral circulation through the oesophageal and perisplenic vascular beds. The fistula was embolized with four Gianturco (occluding string embolus, William Cook Europe, Denmark) coils. On control angiography 15 days later the fistula and its aneurysmic dilatations had disappeared ( Figure Z), and there were fewer arteriographic signs of portal hypertension. Echo-Doppler examination showed hepatopetal portal blood flow. Percutaneous liver biopsy demonstrated chronic active hepatitis and portal fibrosis.When the patient was reviewed 6 months later she was asymptomatic and routine analyses were normal. Endoscopy revealed less prominent oesophageal varices. Digital angiographic studies (April 1990) showed revascularization of the intrahepatic arterioportal fistulous sinus with filling of the portal vein in the early arterial phase and retrograde portal blood flow (Figure 3). Embolization was performed again with five Gianturco coils, almost eliminating the fistula, although some collateral vessels from the right hepatic artery or even from the left hepatic artery (which arose from the left gastric artery) bypassed the obstructive coils and still fed the fistulous area.