Introduction: Because of the current shortage of cadaveric organs, it is important to determine preoperatively those variables that are readily available, inexpensive, and noninvasive that can predict a higher incidence of hepatic artery thrombosis (HAT).
Material and methods: From April 1986 to October 2001, 717 patients underwent 804 liver transplants. All the arterial reconstructions were performed with fine (7-0) monofilament sutures in an interrupted fashion. Two methods were used: group I, end-to-end arterial anastomosis, and group II, the gastroduodenal branch patch.
Results: After a mean follow-up of 72 (range 3-174) months, HAT was observed in 19 patients (overall incidence 2.4%). End-to-end anastomosis (group I) was performed in 39.50% (316) of cases, and HAT developed in 14 (4.4%) cases. Branch-patch anastomoses (group II) were carried out in 60.5% (488) of the patients; the presence of HAT was detected in five cases (1.03%) (P = 0.03, P < 0.05). A total of 21 variables were selected in the univariate analysis; however, after the multivariate analysis, all but two of the factors lost statistical significance, and these corresponded to the type of arterial reconstruction (gastroduodenal branch patch vs. end-to-end) and the ABO compatibility.
Conclusions: Liver transplantation with compatible grafts using branch-patch anastomosis for the arterialization (both manipulative by the transplant team) reduces HAT-derived loss of grafts, with the consequent increase in graft availability and reduced mortality rate on the waiting list.