A study was designed to determine whether soluble mediators of injury are released during cold preservation. A first set of livers consisting of three groups was stored in cold Euro-Collins solution. These were a control group stored for 10 min (group 1), an experimental group stored for 16 hr (group 2), and an "antiprotease" group to which a cocktail of antiproteases had been added, which was also stored for 16 hr (group 3). The preservation solution in these livers was washed out at the end of preservation, and this effluent was concentrated and infused into a second set of livers that were all cold-stored for 4 hr. Then, the second-set livers were either perfused-fixed at 4 degrees C with universal fixative or reperfused at 37 degrees C for 180 min in the isolated perfused rat liver (IPRL). Morphometric assessment of sinusoidal lining cells (SLC) on light and electron microscopy showed an increased degree of microcirculatory injury in livers preserved with concentrates from livers of the experimental group. On light microscopy, only 2.2 +/- 0.4% (mean +/- SD) of the SLC had a normal flattened morphology compared with 11.9 +/- 2.0% in the control group, and 10.7 +/- 2.3% of the SLC appeared completely detached from the underlying hepatocytes compared with 2.6 +/- 0.8% in the control group, the differences being statistically significant (P < 0.05). This injury was prevented by the addition of antiproteases to EC solution. Similar results were obtained in the IPRL model, in which a number of typical changes related to cold preservation injury were noted in livers preserved with concentrates from the experimental group. Compared with controls, livers preserved with concentrates from the experimental group had early and significant alterations in markers of microcirculatory injury, including a reduction in portal flow and an increase in creatinine kinase-BB isoenzyme release, followed by an increase in perfusate transaminases, LDH, and a decrease in bile production. Again the injuries were largely prevented by the addition of antiproteases. There were no differences among groups in the degree of white cell and platelet adherence during reperfusion. Experiments using UW solution showed similar results, indicating that the soluble mediator(s) is not specific for a particular preservation solution. These observations are consistent with the hypothesis that soluble mediators are produced during the hypothermic period, and are responsible for a significant part of cold preservation injury, and that proteolytic reactions are involved in this type of injury.