Current issues of ACP Journal Club are published in Annals of Internal Medicine


Paracentesis and peritoneovenous shunting were equally effective for refractory ascites in cirrhosis

ACP J Club. 1992 Jan-Feb;116:10. doi:10.7326/ACPJC-1992-116-1-010

Source Citation

Ginès P, Arroyo V, Vargas V, et al. Paracentesis with intravenous infusion of albumin as compared with peritoneovenous shunting in cirrhosis with refractory ascites. N Engl J Med. 1991 Sep 19;325:829-35.



To compare the effectiveness of peritoneovenous shunting and paracentesis for refractory ascites in cirrhotic patients.


Randomized, controlled trial.


6 hospitals in Spain.


All patients had cirrhosis, diagnosed either by liver biopsy or by clinical and laboratory findings, and refractory ascites, defined as either no response to a sodium intake < 40 mmol/d and 5 days of treatment with increasing doses of diuretics or ≥ 3 episodes of tense ascites requiring hospitalization in the previous 9 months. Exclusion criteria were serum bilirubin concentration > 171 µmol/L, prothrombin time < 40%, platelet count < 40 × 109/L, serum creatinine > 265 µmol/L, variceal rupture and hemorrhage without sclerotherapy in the last 2 months, hepatocellular carcinoma, and respiratory or cardiac failure. Mean follow-up (± SD) was 437 ± 367 days and 441 ± 414 days for the paracentesis and shunt groups, respectively.


After 5 days on a low-sodium diet without diuretic therapy, patients were randomly assigned to repeated paracentesis (n = 41) followed by intravenous albumin, until ascites disappeared, or to peritoneovenous shunting (n = 48). After discharge, patients were examined weekly for 1 month, monthly for the next 2 months, then bimonthly.

Main outcome measures

Survival, duration of initial hospital stay, and rehospitalization.

Main results

3 patients (7%) in the paracentesis group died during initial hospitalization compared with 6 patients (12%) in the shunt group {absolute risk reduction 5%, 95% CI -7% to 19%, P = 0.4}.* The remaining patients in each group were discharged with little or no ascites. The mean length of stay was 11 ± 5 days and 19 ± 9 days for the paracentesis and shunt groups, respectively (P < 0.01). 37 patients in each treatment group were rehospitalized at least once during follow-up. There were more readmissions for all causes and for ascites during follow-up for patients treated with paracentesis (P < 0.001), but no difference existed between the groups for time spent in the treatment of ascites. Patients receiving paracentesis had greater need for diuretics during follow-up (P < 0.01). Survival was similar in both groups.


Paracentesis and peritoneovenous shunting were equally effective in treating refractory ascites in patients with cirrhosis.

Sources of funding: Fondo de Investigaciones Sanitarias de la Seguridad Social and the Fundació Catalana per a l'Estudi de les Malalties del Fetge.

Address for article reprint: Dr. V. Arroyo, Liver Unit, Hospital Clínic i Provincial, Villarroel 170, 08036 Barcelona, Spain.

*Numbers calculated from data in article.


Refractory ascites, an uncommon complication of cirrhosis, is defined as ascites that fails to respond to medical therapy, that is, salt restriction and the administration of diuretic agents. Although treatment options are limited, peritoneovenous shunting and large-volume paracentesis are both superior to medical therapy in this condition (1, 2).

This well-designed trial by Ginès and colleagues compares large-volume paracentesis and intravenous albumin infusion with peritoneovenous shunting in the management of refractory ascites. Most patients (73%) had alcoholic cirrhosis with relatively good hepatic and renal function. The authors observe that shunting was the more effective therapy in the control of ascites, but the benefit of the procedure was limited by the frequency of shunt occlusions (37% of patients). Previous concerns about the safety of peritoneovenous shunting were not confirmed. Although 6% of patients in the surgical group developed symptomatic disseminated intravascular coagulation, the overall incidence of complications and mortality was similar in both groups.

On the basis of this study, physicians should consider peritoneovenous shunts in stable cirrhotic patients with good hepatic and renal function. As shown by Stanley and colleagues in the VA Cooperative Study, severity of illness rather than method of treatment is the more important predictor of outcome (1). Further studies are needed to choose between these 2 treatments in more severe cases of liver disease and to evaluate the role of liver transplantation in this population.

I. David Shocket, MD
Veterans Affairs Medical CenterBoston, Massachusetts, USA

I. David Shocket, MD
Veterans Affairs Medical Center
Boston, Massachusetts, USA


1. Stanley MM, Ochi S, Lee KK, et al. Peritoneovenous shunting as compared with medical treatment in patients with alcoholic cirrhosis and massive ascites. N Engl J Med. 1989;321:1632-8.

2. Ginès P, Arroyo V, Quintero E, et al. Comparison of paracentesis and diuretics in the treatment of cirrhotics with tense ascites: results of a randomized study. Gastroenterology. 1987;93:234-41.