Coagulation data were collected before and after peritoneovenous shunting for intractable ascites in 19 shunting procedures. After insertion of the shunts, changes consistent with disseminated intravascular coagulation developed in all cases in which good flow of ascitic fluid was obtained. In cases with temporary shunt function, the coagulation variables suggestive of disseminated intravascular coagulation returned toward normal when the flow of ascitic fluid ceased. A fall in the level of fibrinogen degradation products indicated that the shunt had clotted. Bleeding attributable to disseminated intravascular coagulation alone was uncommon. Clotting of the shunts was frequent. The use of heparin improved some of the coagulation variables but did not prevent shunt clotting or clinical bleeding. We conclude that the peritoneovenous shunt induces a moderate disseminated intravascular coagulation and that measurement of fibrinogen degradation products is useful in assessing shunt function.