Hepatic artery thrombosis (HAT) is the most severe vascular complication occurring after liver transplantation, with an incidence ranging from 2 to 9% in adults. Although the ideal arterial reconstruction is often described as a short and non-redundant anastomosis fashioned between the recipient and donor hepatic arteries, there is no strong evidence about this ideal reconstruction in the literature. The aim of this study was to assess the impact of the type of arterial reconstruction on early HAT after primary liver transplantation. We retrospectively reviewed a contemporary MELD era cohort of 282 patients who underwent deceased donor primary liver transplantation from 2007 to 2012. Graft artery was classified as "short" when the section was located at the proper/common hepatic artery or "long" when the celiac trunk was used for anastomosis. Recipient arterial sites for arterial anastomosis were classified in three sites: (1) "distal" (proper hepatic artery or common hepatic artery/gastro-duodenal bifurcation), (2) "intermediate" (common hepatic artery) and (3) "proximal" (celiac trunk-splenic artery-aorta). We used univariate and multivariate analyses to assess the impact of different types of arterial reconstruction on early HAT. Of 282 primary liver transplantations, 17 patients (6%) developed early HAT. Patients with and without early HAT had comparable demographic and operative data. The main anastomotic combination was short graft artery on the recipient-common hepatic artery (n=111, 39%). A long graft artery was used in 91 patients (32%) and was associated with hepatic artery variations (56%; n=51; p=0.001). Arterial reconstructions using a long graft artery (p=0.003), a recipient proximal site as celiac trunk-splenic artery-aorta (p=0.02) and the combination of a long graft artery on the recipient distal hepatic artery (p=0.02) were significantly associated with early HAT. The early HAT rate in patients with a long graft artery was not significantly different between patients with or without donor arterial variation (respectively, 12% (n=6/51) vs. 12% (n=5/40); p=1). In multivariate analysis, the use of a long graft artery, whatever the recipient anastomosis site, was an independent risk factor of early HAT (OR 3.2; 95% CI 1.2-9; p=0.02). The type of arterial reconstruction used for arterial anastomosis during primary liver transplantation has an impact on the occurrence of early HAT. The use of a long graft artery is an independent risk factor of early HAT. Thereby, we recommend the use of a short graft artery with a direct path when feasible to reduce the occurrence of early HAT after primary liver transplantation.