Intractable ascites, refractory to medical therapy, occurs in approximately 10% of patients with ascites from cirrhosis 
and is almost always associated with a grave prognosis. The role of peritoneovenous Denver shunt in control of nonmalignant hepatic refractory ascites was assessed by clinical, anthropometeric, Doppler ultrasound and biochemical means. 
The study was performed on twenty patients with ascites not responding to 400 mg spironolactone and 80 mg furosemide daily, with no bleeding attacks in last three months, serum bilirubin less than 4 mg%, compensated heart as well as renal functions and normal serum amylase level. Ascitic fluid sample revealed total leucocytic count less than 250/ml, no growth in culture, protein content less than 4.5 gm% and negative cytology. 
These patients underwent peritoneovenous shunt under local anesthesia. Post-operatively, they were assessed by 
abdominal girth, body weight, and fluid balance. Hemoglobin, packed cell volume, platelets count and coagulation profile. 
Serum electrolytes, liver functions tests, total proteins and serum albumin. All these parameters were measured daily for two weeks and weekly for two months and monthly through out the study. Hepatic and renal duplex Doppler ultrasonography was carried in all patients before and 4 weeks after shunt operation as well as in twenty healthy controls. These data was analyzed statistically using T-test to compare between the ascitic and control groups while Paired t-test to compare pre and postoperative data. Operative mortality included all deaths within 30 days of surgery. 
Regression of the tense ascites, improvement of the quality of life, improvement of the milieu interior and nutritional 
status were achieved in all patients with no operative related mortality. The mean weight and abdominal girth decreased 
significantly after shunt insertion (p<0.001). Haemoconcentration, urinary output and pulse pressure were markedly 
improved. Significant laboratory alterations in coagulation parameters (p<0.001) consistent with DIC were present in 
virtually all patients not associated with clinically evident DIC. Hospital stay was short (7-10) days. The mean resistive 
index of renal artery showed statistically significant reduction as compared to the pre-operative value (0.78 ±0.32 vs. 0.64 ±0.14 respectively P<0.05). 
Few complications were described but they did not influence the general results. Complications related to shunt insertion were easily prevented and properly managed. Shunt occlusion occurred in six patients (30%) (Peritoneal catheter occlusion 20% and pump chamber occlusion 10%). Gastrointestinal bleeding occurred in four patients (20%) (Gastric erosions 10% - variceal bleeding 10%). Minor complications were observed as improper positioning of either the venous catheter in one patient (left innominate vein) or the peritoneal catheter in two patients (the supracolic compartment or subphrenic space). 
Also one patient had a small subcutaneous fluid collection after shunt obstruction. 
In conclusion, Insertion of the peritoneo-venous Denver shunt seemed to be a minor operative procedure done under local anesthesia with minimal surgical stress. It provided good palliation for all patients with little morbidity and no operative related mortality. It improved renal haemodynamic as indicated by reduction of resistive index of the renal artery. Preoperative injection sclerotherapy as well as proton pump inhibitor prevent post-operative gastro-intestinal bleeding. 
Peritoneo-venous Denver shunt in association with chest tube drainage and pleurodesis 3 weeks after shunt insertion 
784 Egyptian Journal of Surgery succeeded in controlling recurrent pleural effusions secondary to liver cirrhosis. So proper patient selection and careful surgical procedure seems to be mandatory for better results.