Type of donor aortic preservation solution and not cold ischemia time is a major determinant of biliary strictures after liver transplantation

Abstract

The development of biliary strictures (BSs) after liver transplantation (LT) continues to affect 10% to 30% of patients, causing substantial morbidity. The cause of BSs is multifactorial, including technical, immune, and, in particular, ischemic factors. The importance of adequate flushing of the peribiliary arterial tree has been stressed. We hypothesized that high-viscosity (HV) preservation solutions in the donor do not completely flush the small donor peribiliary plexus, leading to inadequate preservation of the bile ducts and posttransplant BSs. To test this hypothesis, we retrospectively compared the incidence of BSs in 2 groups of adults undergoing LT using different types of aortic preservation solution in the donor: group 1 (n = 24), low-viscosity (LV) Marshall solution; and group 2 (n = 27), HV University of Wisconsin (UW) solution. All donors in both groups received additional portal flushes with UW. All LTs were performed between November 1995 and August 1998 at 2 centers by the same surgeon, eliminating a technical bias. Terminal duct-to-duct anastomosis was performed in all recipients except 1 patient in group 1, who underwent a bile duct-to-jejunum anastomosis. BSs were first suspected on clinical and biochemical grounds and then confirmed by endoscopic retrograde cholangiopancreatography. Identical medical protocols were used in all patients. One-year patient survival rates in groups 1 and 2 were 92% and 100%, respectively (P =.9). One-year graft survival was identical to patient survival. The incidence of BSs in group 1 was 4.1% (1 of 24 patients), compared to 29.7% in group 2 (8 of 27 patients; P =.02). The BS in group 1 occurred 4 months post-LT and was anastomotic. BSs in group 2 occurred between 1 and 12 months post-LT and were anastomotic, extrahepatic, intrahepatic, or combined intrahepatic and extrahepatic. There were no significant differences in the following factors between groups 1 and 2: donor age, local versus imported liver, split-liver or full-liver transplantation, incidence of multiple vessels in the donor liver, indications for LT, recipient age, T-tube versus no T-tube, post-LT peak aspartate aminotransferase level, and treatment for rejection. There was no hepatic artery thrombosis or primary nonfunction in either group. Interestingly, cold ischemia time (CIT) was longer in group 1, which had the least incidence of BSs (692 +/- 190 v 535 +/- 129 minutes in group 2; P =.001). Follow-up was longer in group 1 (28.9 +/- 8.3 v 15.6 +/- 8 months in group 2; P =.0001). Preservation costs per procurement were 1.9 times greater in the UW group than in the Marshall group. Donor aortic flushing with an HV preservation solution leads to more frequent BSs compared with an LV preservation solution. The impact of preservation solution outweighs the previously described deleterious impact of prolonged CIT. Mixed preservation solution (Marshall solution in the aorta, UW solution in the portal vein) might protect against BS formation while providing optimal liver graft preservation, function, and survival despite a mean CIT longer than 10 hours.