Ascitic fluid albumin and water flows in patients with alcoholic cirrhosis: effects of peritoneovenous shunting.

Ascites labeled albumin and water kinetics were studied at steady state by intercompartmental clearances in cirrhotic patients whose ascites volumes ranged from 6.4 to 25.2 L. In 20 patients mean (+/- SD) ascitic fluid albumin clearance (equivalent to lymphatic absorption) was 3.5 +/- 1.9 L/day. In seven of them, lymphatic absorption was 3.4 +/- 0.8 L/day, total water absorption from ascitic fluid (outflow equals inflow at steady state) was 125.8 +/- 21.5 L/day. Vascular capillary water absorption (total minus lymphatic) was 122.4 +/- 21.4 L per 24 hours. Lymphatic absorption was less than 4% of total but accounted for all albumin absorbed (13.2 +/- 6.2 gm/day). Maximum lymphatic absorption was less than 10 L per 24 hours; peritoneovenous shunting augmented this function. Mean inflow water albumin concentration was 0.11 +/- 0.06 gm/L. If plasma albumin were diluted to concentration in ascites, calculated water inflow required would be 2.82 L/day. Actual inflow was 47 times that required. The major volume outflow path was through peritoneal venous capillaries; water and small-radius solute absorption through this route increased inflow water albumin concentration to ascites level. Concurrently, whole ascitic fluid was absorbed without sieving through lymphatics. Observed ascites albumin concentration in subjects with advanced cirrhosis was produced by water absorption from a large volume of dilute solution rather than dilution of a small inflow volume, in which albumin concentration originally was hyperoncotic to ascitic fluid. Large-volume transperitoneal water diversion from sources in high-pressure, extrahepatic splanchnic capillaries to absorption through parietal nonportal, low-pressure, peritoneal venous capillaries would significantly reduce portal plasma flow into liver and have deleterious systemic hemodynamic consequences.