Therapeutic results of Denver percutaneous peritoneovenous shunt in cancer patients with malignant ascites.

BACKGROUND: Intractable ascites secondary to malignant disease deteriorates patients' quality of life. The purpose of this study was to evaluate the safety and efficacy of percutaneous peritoneovenous (Denver) shunt in treating intractable malignant ascites in cancer patients. MATERIALS AND METHODS: Thirty-five patients who had undergone Denver peritoneovenous shunt for the treatment of ascites associated with malignant tumor from October 2014 to 2017 were retrospectively analyzed. The demographic characteristics, laboratory values, and complications were recorded. Univariate and multivariate logistic regression analyses were performed. RESULTS: The sites of primary tumor were pancreatic cancer in 19 patients, bile duct cancer in 8, gallbladder cancer in 5, breast cancer in 2, and peritoneal malignant mesothelioma in 1. Palliation of abdominal distention was achieved in 29 patients (82.9%). Postoperative complications of Grade 2 or higher were seen in 11 patients (31.4%), and Grade 5 complications were observed in three patients (8.6%). Patients with a high American Society of Anesthesiologists (ASA) grade and high ascites drainage volume had a significantly higher incidence of postoperative complications than a low ASA grade and low ascites drainage volume, and a multivariate logistic analysis showed that the intraoperative ascites drainage volume was an independent risk factor for all complications. CONCLUSIONS: The Denver shunt for malignant ascites is useful for improving patients' quality of life if the indications are selected properly. Drainage of intraoperative ascites was a risk factor for postoperative complications after the Denver shunt technique in cancer patients with malignant ascites. Further experience and discussion are necessary to establish the patient selection criteria.

Predictive Factors for Treatment Failure after Peritoneovenous Shunt for Hepatic Ascites.

Peritoneovenous shunt (PVS) is used to treat refractory ascites. Here, we identify predictive factors for inhospital death after PVS placement. Thirty-five patients with refractory ascites related to liver cirrhosis and/or hepatocellular carcinoma (HCC) who underwent PVS placement between February 2005 and February 2013 were included in the study. Group A comprised 13 patients for whom the PVS placement outcome was inhospital death. Group B comprised 22 patients who were discharged after PVS placement without complications. Patient background and laboratory data were analyzed to identify risk factors for inhospital death. HCC prevalence in Groups A and B was 92 and 55 per cent, respectively (P = 0.02) and that of portal venous tumor thrombus (PVTT) was 54 and 9 per cent, respectively (P = 0.003). The mean des-γ-carboxy prothrombin (DCP) level in both groups was 15,553 ± 49,330 and 787 ± 2600 mAU/mL, respectively (P = 0.009). Multivariate analysis revealed that the presence of PVTT was the only independent predictor of inhospital death (P = 0.007). The presence of PVTT, HCC, and elevated des-γ-carboxy prothrombin levels are predictors of inhospital death after PVS placement. Therefore, PVS should not be used to treat refractory ascites in patients with these predictors, particularly with PVTT.

Prognosis assessment of patients with refractory ascites treated with a peritoneovenous shunt.

BACKGROUND/AIMS: We investigated the postoperative outcome and risk factors for DIC and mortality in cases of implanted PVS. METHODOLOGY: We reviewed the cases of 65 patients implanted with PVS from 2000 to 2010. Of these patients, 32 were diagnosed with peritonitis carcinomatosa, 21 had liver cirrhosis with hepatocellular carcinoma (HCC), and 12 had liver cirrhosis without HCC. RESULTS: The postoperative morbidity rate was 18.8%, 76.2%, and 58.3% in cases of peritonitis carcinomatosa, liver cirrhosis with HCC, and liver cirrhosis without HCC, respectively. Early death (within 7 days of surgery) was 7.7% (5/65), and the cause of death in all cases was DIC. Underlying disease, low platelet count, prolongation of prothrombin time (PT), and hyperbilirubinemia were the risk factors for development of DIC, whereas underlying disease, prolongation of PT, hypoalbuminemia, and hyperbilirubinemia were risk factors for early death. Multivariate analysis showed that liver cirrhosis with HCC and prolonged PT were the risk factors for DIC. CONCLUSIONS: Patients with refractory ascites due to liver cirrhosis with HCC and those with prolonged PT should not be considered for PVS.

Denver peritoneovenous shunts for the management of malignant ascites: a review of the literature in the post LeVeen Era.

Most case series describing peritoneovenous (PV) shunts for malignant ascites include both LeVeen and Denver shunts. Conclusions based on these studies are no longer clinically relevant since the LeVeen shunt has been discontinued. The purpose of this study was to identify outcomes specific to Denver shunts to establish expected results in the modern era. Case series describing PV shunts for malignant ascites between 1980 and 2008 were identified through a keyword PUBMED search. Whenever possible, results attributable to Denver shunts were abstracted and analyzed. Nineteen series describing 341 patients undergoing 353 Denver PV shunts for malignant ascites were identified. The primary indications for PV shunts were unspecified or cancers of unknown origin (40%), ovarian cancer (16%), and pancreatic cancer (8%). Primary patency averaged 87 ± 57 days. Seventy-four per cent of patients died with functioning shunts. Complications occurred in 38% of patients including occlusion (24%) and disseminated intravascular coagulation (9%). Average survival of all patients was 3.0 ± 1.7 months and shunts provided effective palliation in 75.3%. One and a half per cent of 133 patients who had autopsies were reported to have hematologic dissemination. These results are not statistically different than overall results reported for both shunts combined or LeVeen shunts alone. Studies that report combined outcomes with Denver and LeVeen shunts for malignant ascites are neither negatively, nor positively influenced by one specific shunt. Expectations following PV shunting for malignant ascites do not have to be revised because LeVeen shunts are no longer available.

Percutaneous peritoneovenous shunt for treatment of refractory ascites.

PURPOSE: To evaluate the usefulness of a percutaneously placed peritoneovenous shunt (PVS) in patients with refractory ascites. MATERIALS AND METHODS: Under fluoroscopic and ultrasonographic (US) guidance, the authors placed a PVS in 55 patients (39 men and 16 women; mean age, 56 years) with refractory ascites and symptomatic abdominal distention. The cause of ascites was liver cirrhosis (n = 36), carcinomatosis (n = 17), ruptured cysts with polycystic kidney disease (n = 1), and idiopathic refractory ascites (n = 1). The authors retrospectively evaluated technical feasibility, shunt patency, complications, and clinical outcomes of each patient. RESULTS: The technical success rate was 100%, and symptomatic improvement was achieved in all but one patient. Complications occurred in 17 of the 55 patients (31%): five patients had variceal bleeding; three patients had ascites leakage; two patients each had disseminated intravascular coagulopathy, transient abdominal pain, shunt infection, and venous thrombosis; and one patient had pulmonary thromboembolism. Thirty patients (54%) died 2-690 days after the procedure (mean, 117 days), and their lifetime shunt patency was 84%. Eight patients were lost to follow-up. Seventeen patients were alive for 60-1,200 days, and their shunt patency was 71%. There was no significant difference in shunt patency between the two groups with benign and malignant ascites. CONCLUSIONS: The percutaneous placement of a PVS was a technically feasible and effective method for symptomatic relief of refractory ascites.

Peritoneovenous shunts in patients with intractable ascites: palliation at what price?

Intractable ascites carries great morbidity by affecting appetite, mobility, and quality of life. Peritoneovenous shunts (PVSs) are utilized to abate intractable ascites, although long-term efficacy is unestablished. Thirty male and 18 female cirrhotics, 55 +/- 12 (standard deviation) years of age, failed multiple large-volume paracenteses and diuretic therapy before undergoing PVS. Data were collected until death or the present time. Nine patients (19%) are alive and palliated, four with working shunts [average follow-up (ave. f/u), 30 months] and five without shunts (ave. f/u, 19 months). Thirty-two (67%) patients died: 18 palliated with functional shunts (survival time, 4.4 +/- 5.7 months), 8 unpalliated with dysfunctional shunts (ave. f/u, 3.9 +/- 4.5 months), 4 unpalliated with shunts removed (ave. f/u 5.5 +/- 4.7 months), and 2 with unknown shunt function at death. Function was lost to occlusion in 26 patients, infection in 9, and ligation for disseminated intravascular coagulation in 3. Thirteen patients underwent 18 shunt replacements. At death/present time, 22 (46%) patients were palliated with functioning shunts. Seven patients were lost to follow-up. PVSs provide palliation for intractable ascites short term, but commonly occlude within 1 year. Despite palliation, complications with PVSs are high, and survival is limited.

Peritoneovenous shunts for palliation of the patient with malignant ascites.

METHODS: This study evaluated the effectiveness and morbidity associated with 116 peritoneovenous shunts inserted into 89 patients for symptomatic malignant ascites between 1981 and 1991. All patients had previously failed standard medical therapy of salt and fluid restriction, diuretics, and intermittent peritoneal aspirations. RESULTS: Symptoms from the ascites included discomfort, impaired mobility, early satiety, and dyspnea. The study population was composed of 31 men and 58 women whose ages ranged from 24 to 81 years (mean 56). Primary tumors originated from the following sites: breast, 17; ovary, 17; pancreas, 13; gastrointestinal or related organs, 21; and other sites, 21. Mean shunt patency was 83 days; however, only 31% of the patients studied maintained a patent shunt and lived for > 2 months. Symptomatic relief was achieved in only 57 patients (62%). Forty-four patients (49%) developed one or more complications related to shunt placement. The deaths of 12 patients (13%) within 30 days of surgery were directly related to a complication of shunt placement. Overall, 30- and 60-day mortality rates were 43% and 61%, respectively. CONCLUSIONS: Due to the brief and only fair relief of symptoms with a high associated morbidity and mortality, we believe peritoneovenous shunts are of very limited usefulness in the palliation of malignant ascites and suggest that alternate methods should be considered to manage these patients.

Peritoneovenous shunting of intractable ascites in patients with cirrhosis: improving results and predictive factors of failure.

BACKGROUND: The clinical results of peritoneovenous shunting have not been entirely satisfactory in spite of dramatic improvement of renal function and decrease of ascites. The purpose of this study is to determine whether certain modifications improved our results. METHODS: A modified LeVeen shunt was electively inserted in 56 patients who had cirrhosis with intractable ascites. In 24 patients (42.9%) the Child-Pugh's score was less than 9, and in 32 patients the score was 9 or above. Fourteen patients (25%) had previous variceal bleeding, and 15 patients (26.8%) had previous spontaneous bacterial peritonitis (SBP). LeVeen shunt was modified by the addition of a titanium venous catheter tip. Prophylactic antibiotic therapy was administered to all patients. RESULTS: No operative deaths occurred, and one patient had severe postshunt coagulopathy. Five patients (8.9%) experienced recurrent ascites resulting from blockage from the shunt. The cumulative rate of shunt blockage was 5.6% at 1 year and 12% at 2 years. Seventeen patients (30.3%) have recently had variceal bleeding. The cumulative risk of variceal bleeding was significantly higher in patients with a previous hemorrhage than in those patients without (p < 0.05). Eight patients (14.3%) had SBP after operation. The cumulative risk of SBP was higher in patients with a history of SBP than in those patients without, although the difference was not significant. Fifty-five percent of late deaths were related to variceal bleeding or to SBP. Overall cumulative 1- and 2-year survival rates were 67.2% and 55.2%, respectively. It was 82% and 71% in patients without previous variceal bleeding or SBP. CONCLUSIONS: These data suggest that peritoneovenous shunting might be beneficial only in selected patients. Previous variceal bleeding and/or SBP indicate liver transplantation in suitable patients.

Prognostic value of preoperative tests in the surgical treatment of ascites with the implant of Le Veen shunts in cirrhotics

No período de janeiro de 1978 a outubro de 1982, 32 válvulas de Le Veen foram implantadas em 20 pacientes, dos quais 16 eram cirróticos alcoólicos e quatro pós-necróticos. No presente estudo foram correlacionados dados laboratoriais pré-operatórios destes pacientes com sua evoluçäo pós-operatória. Foram comparados os resultados clínicos dos pacientes que sobreviveram mais de 30 dias (13 pacientes = 65%), com os resultados daqueles que morreram dentro do mesmo período (sete pacientes = 35%). Foram realizadas 14 provas laboratoriais visando definir níveis séricos de hematócrito, hemoglobina, uréia, creatina, sódio, potássio, bilirrubina, aminotransferases, fosfatase alcalina, fibrinogênio, gamaglutamiltransferase e atividade de protrombina. Após análises estatísticas, observou-se que seis dos 14 testes realizados podem ser considerados de valor prognóstico, segundo ordem decrescente de importancia: fibrinogênio, fosfatase alcalina, uréia, gamaglutamiltransferase, bilirrubina e atividade de protrombina. Observou-se que todos os sete pacientes que prematuramente faleceram, tinham três ou mais destes alterados, quando comprados com valores padröes. Baseados nestes dados, conclui-se que estes säo fatores importantes em determinar o prognóstico da sobrevida imediata de cirróticos submetidos ao implante da válvula de Le Veen. Conclui-se, também, que quando três ou mias destes fatores estäo alterados, este procedimento cirúrgico dever ser contra-indicado

Prognostic value of preoperative tests in the surgical treatment of ascites with the implant of Le Veen shunts in cirrhotics.

In the period of January 1978 to October 1988, 32 Le Veen shunts (LVS) were implanted in 20 patients, out of which 16 were alcoholic cirrhotics and 4 postnecrotic cirrhotics. In the present study, we correlated preoperative laboratory data of these patients with their postoperative evolution, comparing the clinical results of patients who survived more than 30 days (13 patients = 65%) with the results of those who died within the same period (7 patients = 35%). For that matter, 14 laboratory tests were performed in order to measure the serum levels of hematocrit, hemoglobin, urea, creatinine, sodium, potassium, bilirubin, albumin, AST, ALT, alkaline phosphatase, fibrinogen, gamma GT and prothrombin activity. After statistical analysis, we observed that 6 of the 14 tests performed could be considered of prognostic value in the following decreasing order of importance: fibrinogen, alkaline phosphatase, prothrombin activity, urea, gamma GT and bilirubin. We observed that all the 7 patients who died prematurely presented 3 or more of these levels altered, when compared with standard values. Based on these data, we concluded that serum levels of fibrinogen, alkaline phosphatase, urea, gamma GT, bilirubin and activity of prothrombin proved to be important factors in determining the prognosis of immediate survival in cirrhotic patients who underwent LVS implantation. We also concluded that when 3 or more of these factors are altered, the implant of LVS is contraindicated, whatever clinical criteria for indication and contraindication were taken into account.

Short- and long-term effectiveness, morbidity, and mortality of peritoneovenous shunt inserted to treat massive refractory ascites of nephrogenic origin analysis of 14 cases.

The experience with 14 patients with end-stage renal disease (ESRD), 13 of them maintained on chronic hemodialysis (x 20.4 months +/- 2.9 SEM) and one following successful renal transplantation, underwent placement of a peritoneovenous shunt (PVS) for refractory ascites that had been present before insertion from two to 15 months (x 5.3 +/- 0.8 SEM). A "specific" cause for the ascites could not be identified in any of the 14 patients. The ascites was an exudate in every patient (protein content greater than 3.5 gm/dl). Twelve patients (86%) obtained significant relief of the discomfort and all effects of the ascites, and objective clinical improvement persisted for at least six months. Nine patients (75%) survived one year and six (50%) survived three or more years. Three patients (21%) had recurrence of ascites because of shunt malfunction; however, two of them were successfully treated with placement of a second shunt. Eight (57%) patients have died since the onset of their ascites (x 14.1 months +/- 3.5 SEM); one death was attributable to PVS placement, while the other seven deaths were due to complications of their ESRD. Insertion of a PVS is an effective therapeutic alternative to palliate the discomfort and ill effects of massive nephrogenic ascites that is often refractory to hemodialysis with ultrafiltration.

Long term results (greater than 5 years) in patients with peritoneovenous shunting for intractable ascites: liver function and cancer mortality.

This report is based on twenty-eight (26%) of 107 patients included in a protocol for prospective evaluation of elective peritoneo-venous shunting for intractable ascites in cirrhosis. These patients had no other procedures and survived more than 5 years after the operation. All patients were free of ascites except one in whom it was mild. One patient refused follow-up. Shunt patency was assessed in 23 patients In 14 patients (60.9%), the shunt was obstructed and the superior vena cava was occluded in 5 of them. In 9 patients (39.1%), the shunt was still functioning. No clinical or biological parameters differentiated these two groups of patients. Of the 24 patients who were alcoholics, 2 abstained completely and 20 significantly reduced their drinking habits. In 25 patients, the Pugh's score improved and was A at the time of the study. Seven patients (25.9%) developed a malignant tumor of the oro-pharynx or digestive tract, all other patients were alive and in good health. This study suggests that patients with intractable ascites treated by a peritoneo-venous shunt may survive for a long period. In patients abstaining from heavy drinking, it may function as a therapeutic bridge permitting spontaneous improvement of liver function. The risk of supervening neoplasms suggests that a continuous follow-up of these patients is warranted.

[Peritoneovenous shunting in intractable ascites of cirrhosis. Results of a prospective study on improving prognostic factors]. / Dérivation péritonéo-veineuse dans l'ascite irréductible de la cirrhose. Résultats d'une étude prospective sur les facteurs d'amélioration du pronostic.

A peritoneovenous (LeVeen) shunt was inserted in 39 patients with cirrhosis and intractable ascites. Based on the results of previous experience, the following procedures were performed to improve outcome: 1) intraoperative drainage of most of the ascites; 2) short-term antibiotic prophylaxis by cefotetan; 3) the use of a titanium venous catheter tip. There was no operative mortality. Operative morbidity was minimal. Mean postoperative in-hospital stay was 19 +/- 5 days. Two patients had recurrence of ascites. This resulted from obstruction of the valve in one patient and of occlusion of the venous catheter in the second patient. One-year probability of shunt failure was 5.8 p. 100. Among the long-term complications, variceal bleeding was the most frequent as it occurred in 8 patients and was responsible for death in 6. One-year probability of variceal bleeding was 18.6 p. 100. Overall one-year survival was 68 p. 100, 79 p. 100 in the group of 19 patients with Pugh scores of 8 and less, 58 p. 100 in the group of 20 patients with Pugh scores greater than 8. These results suggest that technically improved peritoneovenous shunting is a low operative risk surgical procedure with high efficiency in the treatment of intractable ascites in cirrhosis.

[Use of the Leveen shunt in the treatment of clinically intractable ascites]. / Uso da válvula de Le Veen no tratamento da ascite clinicamente intratável.

From January 1978 to August 1987, 21 patients received a peritoneovenous shunt using the Le Veen valve (LVV). The indications criteria were the long-term diuretic therapy failure (mean time = 24.4 months) or resistence to medical therapy during hospital internment. The 21 patients underwent 36 surgeries, being 4 valve position review and 11 changes of LVV. The mean age was 51.6 years. Fifteen patients had alcoholic cirrhosis, 3 postnecrotic cirrhosis, one Budd-Chiari syndrome, one mansoni Schistosomiasis, and one malignant ascites. Ten were Child B and 9 Child C patients. Eight patients with history of previous esophageal varices bleeding (EVB) underwent endoscopic sclerotherapy (EE) before LVV implantation. Seven patients died in the early postoperative period (3 Child B and 4 Child C patients). Three patients died due to EVB and the others as consequence of hepatic failure (one), cardiac insufficiency (one), sepsis (one), and bronchopneumonia (one). The mean follow-up was 19.9 months (1-61). Early LVV occlusion occurred in 4 patients and late valve occlusion in others 4 patients. The LVV changes were done at ambulatorial preceeding. Ten patients (47.6%) died in late follow-up and in these cases death was related to the main disease course. It is concluded that: 1) LVV is a useful therapy in patients with intractable ascites, since it is not the terminal manifestations of disease; 2) early mortality is related to liver function and late mortality to main disease course; 3) ascitic patients with EVB should undergo endoscopic sclerotherapy before LVV implantation.

Denver peritoneovenous shunting for malignant or cirrhotic ascites. A prospective consecutive series.

A Denver peritoneovenous (PV) shunt was inserted in 54 consecutive patients for relief of malignant (24 patients) or cirrhotic (30) refractory ascites. The median age of both groups was 58 years, and the most frequent diagnoses were gastrointestinal (15) or ovarian (7) cancers and alcoholic cirrhosis (25). Median survival time was 1.7 and 3.5 months (range, 0.1-15.5 and 0.1-50.5), and the 1-month mortality 42% and 27%, respectively. Postoperative 24-h urinary output increased by 2-31, and the 1-week weight reduction was 8 and 11 kg, respectively, compared with before shunting. Complete shunt failure was encountered early in two patients, due to catheter malposition and clotting. Four more patients experienced transient failure, for an early dysfunction rate of 11%. A shunt-related operative mortality of 6% was caused by pulmonary oedema (two patients) and sepsis (one patient). Shunt malfunction intervened in almost half (6 of 14) of the cancer patients surviving 1 month but was relieved in all but 1. In 3 of 22 cirrhotic 1-month survivors, the Denver shunt had to be removed owing to clotting or sepsis (2 patients) or revised because of blockage. Seven patients with cirrhosis are alive a median of 18 months (range, 2-51) after PV shunt surgery. Side effects were detected in 22 patients (41%): thromboembolism (9 patients), sepsis (7), initially bleeding oesophageal varices (3), DIC syndrome (2), postoperative hepatic coma (2), ascitic leakage (2), and pulmonary oedema (2). Patients with gastrointestinal cancers or severe cardiac disease did not benefit from the procedure. A history of hepatic encephalopathy or a serum bilirubin level above about 100 mumol/l was a bad prognostic sign. We could confirm the reported considerable morbidity and mortality after PV shunting, but also its efficiency in certain cases. Careful patient selection and follow-up study, timing of operation, and adherence to technical details are mandatory to improve the results.

The Denver type for peritoneovenous shunting of malignant ascites.

Peritoneovenous shunts of the Denver type were inserted into 36 patients to control malignant ascites. The Denver system features a compressible pump chamber bearing a pressure sensitive valve. Initially, all the shunts functioned well. The shunt remained open until death in 21 patients, and at the beginning of the analysis, another two patients were still alive with an open shunt. Blockage of the shunt occurred in 13 patients before death. The cumulative survival time for patients after shunt insertion was 129 months and the cumulative shunt functioning time was 92 months. The over-all median survival time after shunt installation was 13 weeks, and calculated actuarially, the median shunt functioning time for long term survivors was 14 weeks. The cytologic state of the ascitic fluid did not make a statistically significant difference to the blockage-free interval (p = 0.99), neither did the type of primary tumor (p = 0.37). Complications were of a minor type. There was no laboratory or clinical evidence of disseminated intravascular coagulation. Tumor spread through the tubing was seen in one of the three autopsies performed. Denver type peritoneovenous shunting appears to provide effective palliation in the majority of patients. It should, however, only be performed as a last resort.

Morbidity and mortality after peritoneovenous shunt surgery for refractory ascites.

A prospective analysis of the morbidity and mortality after peritoneovenous shunting was carried out in 25 patients who had a total of 27 shunts for refractory ascites. Major complications were limited to the patients in whom ascites was secondary to hepatic rather than peritoneal disease. Immediate postoperative complications followed 17 out of the 23 shunts carried out in patients with liver disease and included septicaemia (two), profound hypotension (two), pulmonary oedema (one), and clinically evident disseminated intravascular coagulation (14). Long term morbidity was again limited to the patients with liver disease and included chronic shunt infection (two) and major venous thrombosis (two). Shunt associated mortality was only seen in the patients with liver disease. Despite late shunt blockage in five long term survivors with alcoholic liver disease fluid retention was easily controlled by simple medical means probably because of improved liver function associated with abstinence from alcohol. It is concluded that: (1) patients with hepatic and malignant ascites respond differently to the insertion of a peritoneovenous shunt; (2) Shunt patency should be monitored regularly in patients with liver disease and, because of the potential for septic and thrombotic complications, if blocked the shunt should be removed and; (3) because of the morbidity and mortality of peritoneovenous shunt surgery in patients with liver disease and refractory ascites, an alternative mode of therapy, such as repeated ultrafiltration and reinfusion of ascitic fluid, may be a more effective initial therapeutic approach especially in patients in whom there is a reversible element to their underlying liver disease.

The LeVeen shunt in the elective treatment of intractable ascites in cirrhosis. A prospective study on 140 patients.

One hundred and forty patients with an intractable ascites complicating a chronic liver disease received a peritoneovenous shunt (PVS) using the LeVeen valve. Operative mortality was ten per cent but was 25% in patients with severe liver failure. Intraoperative drainage of ascites sharply decreased postoperative complications and mortality. One-year actuarial survival rate was 81.4%, respectively 77.7%, 61.3%, and 24.7% in patients with good liver function and moderate or severe liver failure. Variceal hemorrhage occurred in 11 patients and late infection in another 11 patients. Thirty-eight patients (30.5%) had recurrence of ascites. This was mostly due to an obstruction on the venous side of the shunt. An elective portacaval shunt had to be done in 23 patients for recurrence of ascites or variceal bleeding. Among the 57 patients still alive at time of writing, 51 were free of ascites. These results suggest that PVS is an efficient operation. This procedure may be largely indicated in the selected and small group of cirrhotic patients with true intractable ascites and moderate or no liver insufficiency.