Transjugular intrahepatic portosystemic shunts (TIPS) are a recent innovation in the management of portal hypertension. In 1992, we had previously described an instance of severe hemolysis associated with this procedure. This study was undertaken to define and quantify the true incidence of TIPS-associated hemolysis and its clinical spectrum, as well as to test the hypothesis that portal decompression by TIPS would ameliorate hypersplenism in patients with portal hypertension. A total of 60 patients undergoing TIPS for prevention of recurrent variceal hemorrhage (n = 40) or refractory ascites (n = 20) were studied. Forty patients with cirrhosis who were followed concurrently served as controls. At entry, both groups were comparable with the exception of increased ascites in the TIPS group. A total of 7 instances of intravascular hemolysis were identified in 60 TIPS patients, whereas none occurred in controls. Of these, 4 patients were asymptomatic and detected on routine laboratory testing. Hemolysis led to a greater than 4-g/dL decrease in hemoglobin in 2 patients, 2- to 3-g/dL decrease in 2 others and a 3- to 4-gm/dL decrease in 1 patient. Two patients were able to compensate for hemolysis and did not develop anemia. In all but 1 case, the findings of hemolysis subsided by 12 to 15 weeks; in 1 patient, orthotopic liver transplantation was associated with resolution of the hemolysis. Overall, no significant changes in white blood cell or platelet counts were observed in patients undergoing TIPS despite adequate portal decompression. We conclude that TIPS-induced hemolysis occurs in approximately 10% of subjects. However, it is self-limited and rarely requires intervention. Potential mechanisms of such hemolysis are discussed. TIPS is also not recommended as a means of improving platelet counts in patients with severe hypersplenism.