Peritoneovenous shunt in treatment of ascites in patients with cirrhosis. A preliminary report with special reference to pathophysiology.

Peritoneovenous shunts (LeVeen type) were implanted in seven patients with cirrhosis complicated by ascites refractory to diuretic treatment. Three patients died of gastrointestinal bleeding and hepatic coma 1 to 7 weeks after the shunt implantation. The patients who died were those with the most severely impaired liver and kidney function. In two of the four surviving patients (observation time, 5-24 months) the shunt was patent during the observation time, and ascites disappeared. In the other two the shunt closed, in one patient repeatedly following several re-implantations. Enhanced urinary sodium excretion was observed in patients with patent shunts. After disappearance of ascites, the splanchnic venous pressures became less deranged. Long-term change in plasma volume or circulating albumin mass could not be detected. A patent shunt increases the drainage from the peritoneal cavity, but detectable increment in the overall lymph drainage was only found in a patient with a very low pre-shunt value. The findings do not support the 'overflow' theory of ascites formation but rather the 'lymph imbalance' theory. For clinical evaluation of peritoneovenous shunting in the treatment of ascites a controlled clinical trial is essential.