A total of twenty six peritoneovenous shuntings were performed in our department between 1978 and 1984, on twelve cirrhotic and ten cancerous patients with intractable ascites, using Pudenz, LeVeen or Denver type shunting tubes. Reduced ascites was noted postoperatively with statistical significance in both the cirrhotic group (p less than 0.05) and the cancerous group (p less than 0.01), of whom eight cirrhotic and seven cancerous patients exhibited an abdominal girth reduced by 9.7 cm (a 10.9 per cent reduction) in two weeks. A larger urinary output was noted in 13 patients, with a mean increase of 587 ml/day compared with the pre-shunting output. These patients lost a mean weight of 6.13 kg in two weeks, while the nonreduced ascites group gained weight. The post-shunting serum albumin level was higher in the reduced ascites group. The mean functioning periods of the shunts were 5.1 and 2 months, respectively for the cirrhotic and cancerous patients whose shunts were patent. The patients' outcome was shown to depend on the disease itself, with the exception of one patient who died of disseminated intravascular coagulopathy following shunting. In comparing several types of shunting systems, the Denver-type was proven as being the most unlikely to plug or malfunction because of its simpler structure.