Leukocyte adhesion may play a central role in the pathogenesis of preservation-reperfusion injury to liver grafts. We previously showed that lymphocyte adhesion to sinusoids is dependent on the length of cold ischemia. In the present study we examined the mechanisms of lymphocyte adherence after harvesting combined with a short and a long preservation time. The effects of lymphocyte adherence on liver function were also examined. Rat livers were stored at 1°C in University of Wisconsin solution for 45 min or 30 hr and then reperfused at 37°C in the isolated perfused rat liver with isogeneic lymphocytes in an asanguineous perfusate. The role of reactive oxygen intermediates was investigated with allopurinol, a vitamin E analog and ascorbate or superoxide dismutase and catalase. For us to determine the role of Kupffer cells, Kupffer cell blockade was produced by gadolinium chloride. Leukotriene B4 effects were examined with the lipooxygenase inhibitor, nordihydroguaiaretic acid. We evaluated the possible presence of mechanical obstruction by studying flow rates and the circulation of red blood cells. We examined the role of adhesion molecules by pretreating lymphocytes with trypsin or neuraminidase and by exposing livers to arabinogalactan. We investigated the effects of lymphocyte adhesion on liver function by comparing perfusate liver enzymes in livers reperfused with and without lymphocytes, with trypsinized lymphocytes and with an increased number of lymphocytes. Allopurinol significantly reduced hypoxanthine degradation, and nordihydroguaiaretic acid inhibited leukotriene B4 release into the perfusate. The ability of gadolinium chloride to inhibit Kupffer cells was shown by colloid carbon uptake. In livers harvested and preserved for 45 min, lymphocytes decreased about 40% during reperfusion. In livers preserved for 30 hr, the reduction was significantly greater (about 80%). Lymphocyte adherence was lessened in livers preserved for 45 min by all three of the reactive oxygen intermediate protectants and by gadolinium chloride. In contrast, neither reactive oxygen intermediate protectants nor gadolinium chloride reduced adherence in livers preserved for 30 hr. Nordihydroguaiaretic acid had no effect in livers preserved for either 45 min or 30 hr. Portal flow in livers preserved for 45 min and 30 hr was similar, suggesting an absence of mechanical obstruction, and this finding was supported by a complete absence of red cell trapping. Trypsinization of lymphocytes and exposure of livers to arabinogalactan significantly lessened lymphocyte adherence in livers preserved for 30 hr but not in those preserved for 45 min. Pretreatment with neuraminidase significantly increased adherence in livers preserved for 30 hr. Finally, the presence of lymphocytes in the circuit significantly increased liver enzymes in livers preserved for 30 hr, unless the cells were trypsinized. The use of an increased number of lymphocytes increased perfusate transaminase levels. In livers preserved for 45 min, the release of liver enzymes was much lower and was not affected by the presence of lymphocytes. The results of this study demonstrate the existence of two distinct mechanisms of adhesion, one operative after harvesting and a short period of preservation and another operative after longer periods. Lymphocytes have deleterious effects on the reperfusion of livers stored for a prolonged period. (HEPATOLOGY 1993;17:131–142.)