The hepatorenal syndrome (HRS) is a terminal complication of severe liver disease associated with a mortality of 80 to 90%. Although the renal functional abnormalities in the HRS suggest prerenal azotemia, volume expansion with saline, albumin or ascitic fluid rarely results in reversal of the HRS because fluid redistributes from the vascular space. Since the peritoneovenous (PV) shunt causes sustained central volume expansion, it has been advocated for the treatment of the HRS. We prospectively compared the PV shunt (N = 10) to Medical Therapy (MED) (N = 10) on renal function and mortality in 20 patients with the HRS associated with alcoholic liver disease. The HRS was diagnosed on the basis of clinical, hemodynamic, and laboratory criteria. The insertion of a PV shunt resulted in an increase in pulmonary capillary wedge pressure (4.2 +/- 1.1 vs. -1.5 +/- 1.0 mm Hg, P less than 0.01) and in cardiac index (0.8 +/- 0.3 vs. -0.2 +/- 0.3 1/min/m2, P less than 0.05). After 48 to 72 hours, weight (+3.1 +/- 1.1 kg) and serum creatinine (3.9 +/- 0.5 to 5.5 +/- 0.7 mg/dl, P less than 0.001) were increased with MED therapy and decreased (weight: -3.7 +/- 0.7 kg; serum creatinine: 3.6 +/- 0.4 to 3.0 +/- 0.5, P less than 0.05) with the PV shunt. Despite improvement in renal function, only one patient with the PV shunt had prolonged survival (210 days). In the remainder, survival was 13.8 +/- 2.2 days compared to 4.1 +/- 0.6 days with MED therapy. We conclude that the PV shunt often stabilizes renal function, but does not prolong life in patients with the HRS.