Rabbit livers were stored cold for periods of 6 or 24 hr and tested using the isolated perfused liver model. Five solutions were tested: Eurocollins (EC), Ross and Marshall's hypertonic citrate (HC), modified plasma protein fraction (Cambridge PPF), Ringer lactate, and the recently developed "University of Wisconsin" (UW) solution. After storage livers were perfused with an erythrocyte-free oxygenated Krebs-Henseleit solution containing 4% bovine serum albumin at 38 degrees C for 2 hr. Bile production proved to be the most sensitive index of liver function for discriminating between the various storage solutions and the different preservation times. After 6 hr of cold storage, bile production was similar to control liver bile production (9.8 +/- 2.4 ml/2 hr/100 g) in livers stored in HC (8.8 +/- 2 ml), PPF (9.9 +/- 2.2 ml), and UW (10.3 +/- 1.9 ml); it was slightly depressed in EC (6.7 +/- 2.5 ml, P = 0.06), and markedly depressed in Ringer lactate (4.3 +/- 0.8 ml, P less than 0.05). After 24 hr of cold storage bile production in UW-stored livers was near normal (9.3 +/- 0.7 ml) but significantly depressed (3.5-6.2 ml) in all other solutions tested. Release of enzymes into the normothermic perfusate was also measured (aspartate aminotransferase, alanine aminotransferase, and lactate dehydrogenase). In this small series the differences between cold storage solutions did not always reach statistical significance although the trend was for less enzyme release in livers stored in UW solution. This technique permits rapid assessment and refinement of new storage methods and new solutions for liver preservation prior to testing in a large animal transplant model. The results suggest that UW solution is superior to other preservation solutions and would permit successful 24-hr storage of livers.