Safety of cardiac catheterization in patients with end-stage liver disease awaiting liver transplantation.

Patients with end-stage liver disease (ESLD) are predisposed to bleeding complications due to thrombocytopenia, reduced synthesis of coagulation factors, and increased fibrinolytic activity. The exact incidence of vascular access site and bleeding complications related to cardiac catheterization in this group remains unknown. Eighty-eight consecutive patients with ESLD who underwent left-sided cardiac catheterization from August 2004 to February 2007 were identified. Eighty-one patients without known liver disease matched for age, gender, and body mass index who underwent left-sided cardiac catheterization during the same period were chosen as the control group. Vascular complications were defined as hematoma >5 cm, pseudoaneurysm, arteriovenous fistula, or retroperitoneal bleeding. Patients with ESLD had lower baseline mean hematocrit (32.3 +/- 6.0% vs 39.2 +/- 6.2%, p <0.001) and mean platelet count (90.1 +/- 66.3 vs 236.1 +/- 77.1 x 10(9)/L, p <0.001) compared with controls. They also had higher mean serum creatinine (1.9 +/- 1.7 vs 1.2 +/- 0.8 mg/dl, p = 0.002) and mean international normalized ratio (1.6 +/- 0.7 vs 1.1 +/- 0.2, p <0.001). There were more complicated pseudoaneurysms in the patients with liver failure (5.7% [5 of 88]), compared with 0% in controls (p = 0.029). Patients with ESLD had lower starting hemoglobin levels and greater reductions in hemoglobin after cardiac catheterization, resulting in greater need for packed red blood cell transfusion (16% vs 4%, p = 0.008), fresh frozen plasma (51.7% vs 1.2%, p <0.001), and platelet transfusions (48.3% vs 1.2%, p <0.001). Major bleeding was higher in the ESLD group (14.8% vs 3.7%, p = 0.014), driven mainly by the need for blood transfusion. In conclusion, despite severe coagulopathy, left-sided cardiac catheterization may be performed safely in this patient population, with correction of coagulopathy and meticulous attention to procedural technique.