Preservation-reperfusion injury of hepatic allografts is thought to be associated with Kupffer cell activation and TNF release from the transplanted organ. Confirmation that the allograft is the source of this TNF in an in vivo model is difficult because of rapid equilibration of this cytokine into all compartments. A novel experimental design was devised to aid in accurate localization of the site of TNF release following a orthotopic liver transplant (OLT). In the first group (anhepatic), livers were removed from rats and splanchnic and systemic venous returns were then reestablished using a conduit of donor IVC and portal vein with a portasystemic shunt. In the second group (asplanchnic), the liver, stomach, pancreas, and intestine of the recipient were removed and a donor liver was reimplanted using the recipient IVC as the source of portal blood. The third (OLT-16) and fourth (OLT 8) groups underwent standard OLT with preservation times of 16 and 8 hr in 4 degrees C Euro-Collins solution, respectively. TNF levels were significantly increased in the OLT-16 group compared with the OLT-8 group. There were modest elevations of TNF in the anhepatic model, but the TNF in the asplanchnic model approached baseline. Absence of TNF in the asplanchnic group and a rise in TNF levels in the anhepatic group to that not significantly different from OLT-16 or OLT-8 suggest that a major source of TNF following preservation reperfusion may be the intestine.